feat: report package link to package document elements

rdeltour · rdeltour · commit dd00b885acf8 · 2022-12-23T18:24:39.000+01:00
This commit introduces a new error message (`OPF-098`) to report `href` attributes
containing a URL to package document elements.

This implements a check for the following specification statement:
&gt; The URL string MUST NOT reference resources via elements in the package document
diff --git a/src/main/java/com/adobe/epubcheck/messages/DefaultSeverities.java b/src/main/java/com/adobe/epubcheck/messages/DefaultSeverities.java
@@ -292,6 +292,7 @@ private void initialize()
     severities.put(MessageId.OPF_096, Severity.ERROR);
     severities.put(MessageId.OPF_096b, Severity.USAGE);
     severities.put(MessageId.OPF_097, Severity.USAGE);
+    severities.put(MessageId.OPF_098, Severity.ERROR);
 
     // PKG
     severities.put(MessageId.PKG_001, Severity.WARNING);
diff --git a/src/main/java/com/adobe/epubcheck/messages/MessageId.java b/src/main/java/com/adobe/epubcheck/messages/MessageId.java
@@ -286,6 +286,7 @@ public enum MessageId implements Comparable<MessageId>
   OPF_096("OPF-096"),
   OPF_096b("OPF-096b"),
   OPF_097("OPF-097"),
+  OPF_098("OPF-098"),
 
   // Messages relating to the entire package
   PKG_001("PKG-001"),
diff --git a/src/main/java/com/adobe/epubcheck/opf/OPFHandler30.java b/src/main/java/com/adobe/epubcheck/opf/OPFHandler30.java
@@ -400,10 +400,20 @@ private void processLink()
     if (url != null)
     {
       // Data URLs are not allowed on `link` elements
-      if ("data".equals(url.scheme())) {
+      if ("data".equals(url.scheme()))
+      {
         report.message(MessageId.RSC_029, location());
         return;
       }
+      // The `href` attribute MUST not reference resources via elements
+      // in the package document itself
+      if (url.fragment() != null && !url.fragment().isEmpty()
+          && URLUtils.docURL(url).equals(context.url))
+      {
+        report.message(MessageId.OPF_098, location(), href);
+        return;
+      }
+
       if (context.isRemote(url))
       {
         report.info(path, FeatureEnum.REFERENCE, href);
diff --git a/src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties b/src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties
@@ -213,7 +213,8 @@ OPF_094=The "media-type" attribute is required for "%1$s" links.
 OPF_095=The "media-type" attribute of "voicing" links must be an audio MIME type, but found "%1$s".
 OPF_096=Non-linear content must be reachable, but found no hyperlink to "%1$s".
 OPF_096b=No hyperlink was found to non-linear document "%1$s", please check that it can be reached from scripted content.
-OPF_097=Resource "%1$s" is listed in the manifest, but no reference to it was found in content documents. 
+OPF_097=Resource "%1$s" is listed in the manifest, but no reference to it was found in content documents.
+OPF_098=The "href" attribute must not reference resources via elements in the package document itself, but found URL "%1$s".
 
 #Package
 PKG_001=Validating the EPUB against version %1$s but detected version %2$s.
diff --git a/src/test/resources/epub3/05-package-document/files/link-to-package-document-id-error.opf b/src/test/resources/epub3/05-package-document/files/link-to-package-document-id-error.opf
@@ -0,0 +1,19 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<package xmlns="http://www.idpf.org/2007/opf" version="3.0" unique-identifier="uid"
+    xmlns:dc="http://purl.org/dc/elements/1.1/"
+    prefix="example: https:example.org">
+    <metadata>
+        <dc:title>Title</dc:title>
+        <dc:language>en</dc:language>
+        <dc:identifier id="uid">NOID</dc:identifier>
+        <meta property="dcterms:modified">2019-01-01T12:00:00Z</meta>
+        <!-- link MUST NOT reference resources via elements in the package document -->
+        <link rel="example:property" href="#t001" media-type="application/xhtml+xml"/>
+    </metadata>
+    <manifest>
+        <item id="t001" href="contents.xhtml" properties="nav" media-type="application/xhtml+xml"/>
+    </manifest>
+    <spine>
+        <itemref idref="t001"/>
+    </spine>
+</package>
diff --git a/src/test/resources/epub3/05-package-document/package-document.feature b/src/test/resources/epub3/05-package-document/package-document.feature
@@ -31,12 +31,11 @@ Feature: EPUB 3 — Package document
   
   @spec @xref:attrdef-href
   Scenario: 'link' target must not reference a manifest ID
-    When checking file 'link-to-publication-resource-error.opf'
-    Then error OPF-067 is reported
+    When checking file 'link-to-package-document-id-error.opf'
+    Then error OPF-098 is reported
     And no other errors or warnings are reported
 
 
-
   ### 5.3.3 The id attribute
   
   Scenario: 'id' attributes can have leading or trailing space 
